Hi, I'm Sohith Bandari (billa-man) 👋
AI Engineer
I build and deploy end-to-end ML systems, from LLMs to scalable production pipelines.
Recent Projects
- MLOps Pipeline for Chest X-Rays (Ray, MLflow, Kubernetes, ArgoCD, Terraform, Triton, Prometheus, Grafana)
- A production-grade MLOps pipeline for medical imaging, enabling distributed training, automated CI/CD deployment, and real-time model serving with monitoring and scalability.
- End-to-End LLM Development with RAG (LangChain, ClearML, MongoDB, Qdrant, Docker, Ollama, AWS)
- A domain-specific RAG system, integrating ETL pipelines with vector search and a fine-tuned Llama-3.2 model for real-time, context-aware query responses.
- Soothify (Next.js, React, Hume EVI, MongoDB, Node.js)
- An AI-powered conversational companion that enables real-time voice and text conversations using Hume EVI and OpenAI.
Research
- Exoplanet Classification through Vision Transformers with Temporal Analysis – The Astronomical Journal, IF 5.1 (2025).
- Thesis (IIT-ISM, 2024): Prediction of Per-Packet Delay in SDNs using Graph Convolutional Networks.